This project addresses statistical problems generated from collaboration with scientists in other program areas and general statistical problems of current interest. This project is a continuing activity of the Section on Mathematical Statistics. Papers have been submitted or published in FY 1988 on the following statistical subjects: estimation of the joint survival and censoring distribution in the presence of dependent censoring; extrapolation of attributable risk to new populations; robust selection procedures using vector ranks; nonparametric statis- tical methods for mutagenesis assays; and casefinding and coverage as statistical methodology issues in population-based area studies. Other work in progress includes: selection criteria for use of the Kaplan-Meier or parametric MLE for survival analysis; derivation of inferential methods for determining order of categorical data; influence of missing data in randomized clinical trials; methods to improve coverage in surveys; analysis of time-to-event data with non-regular censoring; discrete state semi-Markov models with time- dependent covariates imbedded in proportional hazard transitions; estimation of time-to-event with interval data in the presence of left and right censoring; influence of left truncation of time- dependent covariates on survival models; comparison of param-etric and nonparametric survival analysis with mismodeling; site selection for epidemiological surveys; adjustments for covariates in the analysis of categorical data; and use of covariates and statistical models for epidemiology studies.